September, 1945
INDUSTRIAL AND ENGINEERING CHEMISTRY
esterification), increased decomposition of the product resulted in marked darkening of the reaction mixture. An esterification was therefore run without sulfuric acid. Table IX and Figure 8 give the results which show that the reaction is slow even at 120" C., and that sulfuric acid catalyst must be used. CONCLUSIONS
1. The mechanism for esterification of diol to diacetttte proreeds via two pairs of consecutive, reversible reactions of approximately equal speed. 2. ' The presence of excess diacetate is disadvantageous to the euterification rehction. 3. An empirical method of correlating kinetic data with temperature of reaction consists of plotting the lo ari;thm of l/e against 1/T. The slopes of the straight lines should mve the energy of activation for the reaction involved. 4. MVCA can be used for an entrainer of water formed. It is preferred because it is obtained during the pyrolysis of the diacetate to butadiene. 5. Ninety-five per cent conversion to the diacetate is obtained on a laboratory column using MVCA as an entrainer to remove the water of reaction. I n the operation of an esterifica-
attained
905
tion column, the temperature in the lower sections of the column can be prevented from going above 120" C. for any extended period by the use of vacuum. ACKNOWLEDGMENT
Appreciation is gratefully expressed to Schenley Research Institute and to A. J. Liebmann for aasistance and supply of materials used in this work. LITERATURE CITED
Glasstone, S., "Textbook of Physical Chemistry", p. 1067, New York, D. Van Nostrsnd Co., 1941. (2) Northern Regional Research Lab., Summary of Research on Production of Butadiene from 2,SButylene Glycol by Pyrolysis of Diaaetate, Aug., 1942. (3) Othmer, D. F., U.S. Patent 2,050,234 (Aug.4, 1936). (4) Othmer, D. F., Bergen, W. S., Shlechter, N., and Bruins, P. F., (1)
Ibid., 37,890 (1946).
(5)
Othmer, D. F., Shlechter, N., and Koszslks, W.,Zbid.,. 37, 895 (1945).
PRESENTED a8 part of the Unit Proceas Symposium before the Division of Industrial and Engineering Chemistry at the 108th Meeting of the AMERX(IAN CHEMICAL SOCIETY in New York, N. Y.
PYROLYSIS of 2,S-BUTYLENE GLYCOL NATHAN SHLECHTER, DONALD F. OTHMER, AND ROBERT BRAND POLYTECHNIC I N S I I T U T ~ Q P BROOKLYN. N. Y .
T h e final stage in t h e production of butadiene by fermentation of carbohydrate materiais is t h e pyrolysis of butylene glycol diacetate t o give butadiene and acetic acid, with t h e ultimate purification of butadiene for synthetic rubber manufacture. The pyrolysis of butylene glycol diacetate was studied, and yields were determined under various temperatures and other operating conditions. Optimum conditions appeared t o be a contact t i m e of 7.1 seconds and a temperature of about 685' C. in an atmosphere of nitrogen. T h e yield a t these conditions was 84.9% conversion t o butadiene. Acetic acid recovery was good; polymerization of butadiene and formation of unsaturated esters during t h e pyrolysis reaction seem t o be t h e most important factor in determining yields, particularly a t higher temperatures and higher contaot times. T h e amount of polymerization was reduced t o a minimum by diluting t h e gases with sufflclent nitrogen t o reduce t h e concentratlon of diacetate in t h e vapors t o about 40 mole %. lnoomplete pyrolysis and formation of unsatdrated esters may be minimized by subsequent passes of t h e condensate liquor following pyrolysis.
HE recovery of 2,3-butyleiie glycol and its esterification to the diacetate have been reported (3,4, 6). The structural formula of this diol indicated that butadiene might be obtained by the removal of two molecules of water. To assist this reaction, many catalysts, selected for dehydration properties, have been tried by other investigators (9). The maximum conversion (20733)is too low for commercial application. Hill and Isaacs (I) claim much higher yields of butadiene by the pyrolysia of 2,3-butylene glycol diacetate. This pyrolysis
T
probably proceeds by way of an unxaturated ester according to the following equation:
-
CHx ?Ha bHOOCCH8 -CHaCOOH bH OHOOCCHI
bHI 2,3-B1itylene BlYOOl
diaoetata
heat
B%
PHs
bH
-CHaCOOH 6 H ____c
~ O O C C H -t I ~HOOCCHI
bH8 Acetate of enolio methyl ethyl ketone
bH, rMeth 1 vinyl oar&nol aoetate
heat
8H
bH 1, a-Butadiene
No butylenes are formed, although incomplete cracking yields unsaturated esters which boil around 118" C. Butylenes form during the pyrolysis of other materials t o give butadiene; since they boil (-6" C.) very close to butadiene (-5" C.), they are separated from butadiene only with great difficulty. Original work on this pyrolytic reaction, particularly that a t the Northern Regional Research Laboratory, has shown that temperature in the reaction tube and rate of feed are the important factors in determining the extent to which 2,sbutylene glycol diaoetate is converted to butadiene and to by-products. Hill and Isaacs (1) indicate that 475" C. i s about the lowest limit for butadiene production; above 000' C. decomposjtion of butadiene is serious. At the lower temperatures, evidently, unsaturated esters are formed. T o crack these unsaturated compounds further to butadiene, higher temperatures are necessary. Since butadiene polymerization begins to take place at the higher temperatures, some optimum temperature for the operation of the pyrolytic reaction must be found. Northern Regional Research Laboratory has also demonstrated that the slower the rate of feed a t a given temperature, the greater the amount of polymerization which takes place, whereas the unsaturated ester formation decreases with slower feed rates.
906
INDUSTRIAL A N D ENGINEERING CHEMISTRY
Vel. 37, No. 9
F was 2 feet long with a %inch inside-diameter bore. The heating element was a heavy-duty Nichrome resistance coil (1500 watts capacity) wound around a ceramic core and then covered with a 2-inch layer of 85% magnesialaggin Reaction chamber J was a 47-mm. unpacked gyrex tube, 33 inches It was inserted into the vertical furnace. Ion Wize-mouthed thermos bottles H , containing a saturated solution of solid carbon dioxide and acetone, where used as butadiene condensers, The vaporizer was calibrated for the relation between electrical power (watts) input and rate of vaporization (grams per minute). Thus ammeter and voltmeter readings indicated the amount of vapor fed to the cracking unit by use of a calibration chart. Experiments were run at several temperatures to find the effective reaction zone for estimating contact times. This zone was considered that portion included in a range 25' C. below the maximum, and was assumed equal to 7 inches in length. During a run the thermocouple waa inserted to a position 16 inches from the bottom which is at the maximum temperature region. The temperature gradient across the diameter of the tube was negligible. OPERATION.Furnace F was brought to the desired temperature. Vaporizer A was charged Figure 1. Apparatus for Pyrolysis Experiments with a weighed amount of butylene glycol diacetate; and the desired feed rate was set. Cold A. Vaporizer J . Reaction chamber traps H I and Hzwere cooled to about -75" C. E . Heater f o r vaporlzer L. Carbontrap C. Three-way stopcocks R. Llquid take-off I w e i v e r s When the thermometer in the vaporizer reached D . Nitrogen inlet S. Water scrubber the boiling point of the diacetate (191' C.). cock E. Superheater T. Glass tublng F. Furnace TH. Thermometer or Cl was tuined to admit vapors into the reaction C. Calcium chiorlde drier thermocouple tube. A hot Nichrome wire was wrapped around H. Cold traps CON. Condenser I . Butadiene receivers RC. Rubber connections glass tubing T to prevent condensation. The vapors were heated to about 430" C . by superheater E . The moment cock C1was opened, an electric timer w&s started. The reaction was permitted to reach steady state The commercial possibilities of this process depend largely on conditions by running for about 5 minutes. At the end of this the efficiency of the cracking operation and the attainment of period cock C, was turned to permit the products to enter the maximum conversion. From theoretical considerations, it was main receiver, R, and the time was noted. concluded that the use of a hot inert gas (nitrogen) might be At the end of a run the condensed liquid take-off was careful1 heated to the boilin point to remove dissolved butadiene, whici advantageous and useful in obtaining higher yields. Since was subsequently coflected in a cold trap. The products collected nitrogen is a diluent, it reduces the partial pressure of the resulting in R1, I,, I;, and L were weighed. butahene and thus tends to retard the polymerization reaction, Three independent checks were obtained on the feed rate of The presence of air or any oxidizing material g:eatly facilitates vapors to the reaction chamber: ( a ) The voltmeter and ammeter readings and the calibration chart showed the amount vaporized. the polymerization of butadiene ( 6 ) . (6) A check rate was run either before or after the actual run, when cock C, was so turned that the vapors were going to conPYROLYSIS APPARATUS denser I. The condensed vapors were collected in a tared receiver over a measured time interval and weighed. ( c ) An overFigure 1 is a diagram of the apparatus used for the experiall material balance was obtained with the original charge to the mental work. Vaporizer A consisted of a wide-mouthed vacuum vaporizer. The residue a t the end of the run and the material thermos bottle covered with a 1-inch layer of 85% magnesia collected in receivers beneath condenser I were measured. The lagging. Heater for the vaporizer was provided by Nichrome difference was the input to the reaction chamber during the run. wire B (rated at 600 watts), inserted into a Pyrex tubing bent Analyses were also made on the liquid take-off and the butainto a U. Current for the vaporizer was supplied from the 110diene fractions. The take-off liquors from the pyrolysis convolt d.c. line and was regulated to control the rate of vaporization. tained acetic .acid, unsaturated acetates, hydrocarbons (polyA chromel-alumel thermocouple was used t o read furnace temmers of butadiene), unreacted diacetate, and dissolved butadiene. ratures. It was inserted in a silica tube to insulate it from the The following analytical method was used to determine the ot vapors. amount of the various materials present. It is based mainly on Superheater E was an electrical heater which extended for a s method worked out at the Northern Regional Research Laboraheight of 8 inches into the reaction chamber. the heat input wm controlled by a Variac transformer. Electriklly heated furnace tory (3).
.
.
B"
T.4BLE
I.
PYROLYSIS OF
DIACETATE WITHOUT NITROGEN
Feed Liquid Take-off Rate, Contact DiRCetRte Temp., Grams/ Time, Input, Total, Unsatd* C. Min. 6ec. Grams grams Grama %a 'Grama %a 11.2 5.8 12.4 500 5.20 9.2 210.7 169.4 23.8 6.2 8.9 4.6 202.1 154.5 16.1 500 3.20 15.0 0.76 0.74 11.1 10.8 600 0.87 54.3 103.9 77.6 27.2 104.1 79.0 3.43 3.38 7.7 7.6 600 1.70 9.9 90.1 4.52 3.78 7.1 6.0 550 4.50 121.3 4.52 7.3 6.3 87.8 5.14 550 5.80 7.7 116.2 0.76 0.82 12.9 13.8 650 0.77 57.8 94.8 73.6 2.36 2.50 11.8 10.3 650 1.95 22.9 116.6 90.1 685 2.70 15.8 121.1 91.8 3.61 3.02 10.7 8.9 585 1.85 23.2 164.8 127.3 2.85 1.73 19.3 11.8 585 1.05 40.7 167.0 127.6 1.71 1.03 23.3 14.0 585 6.08 8.5 160.9 118.7 5.75 3.62 13.5 8.5 a Weight per cent of diester reacted. b Weight per cent of diester input..
Unreacted Diwter Grams % b 19.2 9.2 6.6 3.3 1.7 1.6 2.7 2.6 2.0 1.6 2.4 2.1 1.5 1.6 1.9 1.6 1.2 1.0 0.8 0.5 Ned. 0 2.1 1.3
Acetic Aaid Tot4 Free input, So reGram; g r a m covery 115.2 145.0 Y6.9 122.9 139.0 97.6 71.7 91.5 64.0 72.0 95.6 65.2 76.5 83.7 95.7 72.8 80.0 95.1 58.2 65.4 90.3 74.4 80.5 96.0 76.3 83.7 94.0 104.3 113.5 93.7 102.6 115.0 90.0 97.3 110.0 91.7
70con- Recovered, Grams version Wt. % 37.1 62.5 98.0 44.5 73.4 98.5 20.2 63.5 94.4 22.1 70.4 96.7 27.9 75.3 97.2 97.4 25.2 71.5 15.6 54.0 94.0 98.0 24.0 66.7 24.4 65.7 96.1 30.2 59.4 95.6 27.9 63.8 93.1 37.4 76.0 97.3
INDUSTRIAL AND ENGINEERING CHEMISTRY
September, 194s
901
TABLE 11. PYROLYSIS OF DIACETATE IN PRE~SE~NCEI OF NITROQEN Flow Rates Dieater NI Contaot
Tpng., gram/ o?.) min.
586 585 585 550 650 550 0
nun.
Time, aeo. 8.5 5.2 7.1 4.6 8.1 9.7
1.1 3.8 2.1 5.4 1.3 0.79
Diaaetate Input Mole
Totsl in 7 % %
Grams mixt. 70.7 144 7 126.5 221.0 94.6 87.4 I
22 46 35 56 25 17
Liquid Take-oB
grams Grams 56.7 107.3 .92.6 177.0 71.4 66.7
8.2 14.0 11.5 41.5 14.9 12.5
%a
8.2 9.9 9.2 20.3 16.6 14.7
Grama 2.7 3.1 3.9 2.4 1.4 2.1
G
%a
3.6 2.2 3.1 2.1 1.6 2.5
Unreaoted Diester
1.1 2.8 1.9 16.4 4.0 2.2
Aoetio Aoid
C
Free, gram
~ grams
input,
reaovery
1.4 1.9 1.5 7.4 4.4 2.5
46.7 87.4 75.a 116.7 51.1 49.9
53 0 Q9:6 87.1 152.5 65.3 60.2
96.5 97.1 95.0 98.1 94.8 96.2
Butadiene ~ ~ w O/oReoovered Grams version Wt. % 19.5 34.7 32.5 43.0 22.6 20.3
83.2 78.9 83.9 67.8 79.5 77.1
99.2 98.2 98.9 99.5 99.1 99.6
Weight per oent of diester reacted. of dieater input.
b Weight per cent
The dissolved butadiene was boiled out of the takeoff liquors and collected in a cold trap. A sample of the butadiy-free liquora was analyzed for total free and combined acetic acid; t b provided an over-all check on the anal me. Them li uora were and the then fractionated. One fraction b o i l d below 145' other above. Each portion was weighed and analyzed for free and combined acetic acid. From these analyses, it was possible to calculate the acetic acid, unsaturated acetates molecular weight 114) and unreacted diacetates in the liquid t$i e-off. The combined acetic acid in the dlstillate boiling up to 145' C. was*consid?red.tobe from the unsaturated acetate whle the combmed acid in the fraction b o i above 145' waa considered to be from glycol diacetate. The dflerence between the sum of calculated values of percentage content of free acetic acid plus unsaturated acetates plus unreacted diacetate, and 100% waa considered t o be the percentage hydrocarbon in the take-off liquors.
attempt to approximate the time spent by the reactants in the
reaction 'One:
8.
Where
The lack of exact material balance is attributed to the destruction of some acetic acid at the temperature of the reaction. The recoveries for the runs without nitrogen, however, are good, ranging from 93.1 t o 98.0% and averaging 90.4%. Important also from an economic viewpoint is the recovery of acetic acid which is also good, 90.0-97.0%, with an average of about 94%. Pyrolysis with nitrogen resulted in increased weight recoveries of 96.4 to QQ.l%, and acetic acid recoveries from 94.0 to 96.1%. This is probably due to the retardation of acetic acid destruction at the higher temperatures by the presence of nitrogen. Figures 2 and 3 show that for each temperature there is an optimum rate of feed at which a maximum yield of butadiene ip obtained. At the slow rates lower temperatures result in larger yields of butadiene; at the faster rates, the curves cross and show higher yields. This is in line with theoretical considerations. At higher temperatures and slow rates, butadiene polymerization dominates the reaction; a t lower temperatures and faster rates, incomplete aracking (only to the unsaturated ester) prevails. Figures 4 and 5 show that a t each temperature there is a steady increase in the amount of unreacted diacetate and unsaturated eaters with decreasing contact times (increasing rates). These curves also illustrate that, for a given contact time, the percentage of w e a c t e d diacetate and unsaturated esters decrease with increasing temperatures. I n general, the percentage of unreacted diacetate is much less than the quanti ies of unsaturated esters formed. The data for polymer formation (Figure 0) are not so regular 8s are the rest of the data, since the amount of polymer was ob-
d
"0
A
Figure 2.
.a
1.D
1.6
-
e.0 2.4 2.6 3.2 3.6 4.0 4 4 4.6 A%? 5.6 6.0 RATE GRAYS 1 YIN.
Conversion t o Butadiene
VI.
contact time VR/VO volume of reaction zone VO = volume of incoming stream per unit time at temperature and pressure of reaction zone vg
Rate of Flow
Assumin these hydrocarbons to be polymers of butadiene, it was poeaibfe to calculate the percentage yield of each of the grodupts from the diacetate fed into the pyrolysis system. The u t d e n e waa analyzed by careful fractionation; and a fraction waa collected boiling from -5' to 0' C. The nonvolatile material remaining was recorded. The butadiene analyzed in this manner waa 9 8 4 9 % pure. PYROLYSIS REACTION
A complete study was first made of the pyrolysis reaction at various temperatures and feed rates to obtain information as to butadiene yields, unsaturated ester fonnstion, hydrocarbon production, and acetic acid recoveries (Table I). Then a series of experiments was run at selected temperatures and feed rates to illustrate the effect of cracking in an atmosphere of nitrogen (Table 11). The flow of nitrogen was measured by a calibrated orifice meter. T o compare results obtained by various investigators in equipment of varying sizes and to use laboratory and pilot plant results for plant design, a common expression for the rate of feed must be used. The general term R i contact time, which is rtn
691
60
.
Sb
69
40
Figure 3. Thio plot 10 In -nd
44
40
W 39 PO 94 CONTACT TIYC' ECCONDO
Conversion t o Butadiene
V.I
90
I6
I9
0
4
Contacrt t i m e
quadrant b-ur a hlgh eontaet tlmo is oqulvalont to a low flow rate.
I
D
i
~
908
INDUSTRIAL AND ENGINEERING CHEMISTRY
FIGURE 4
'i 2
n W
t
4
t
3
!
,
,
,
CONTACT TIME -SECONDS
,
.
,
d
FIGURE 5
14-
c p9 11802 9
p
10-
3)
c 8 -
2 2
6 -
a
e60 I5
56
52
48
44
40 36 32 28 24 20 CONTACT TIME SECONDS
-
16
12
8
4
13
II
Vol. 37, No. 9
greatest deterring factor for high butadiene yields. Thus, any scheme of operation which cuts polymer formation will great,ly increase the efficiencyof pyrolysis. Pyrolysis in an atmosphere of nitrogen increased t,he opt,imum convers'on to butadiene (Figures 2 and 3) from 76YGwithout to 84% with nitrogen. Most' important, however, is the reduced amounts of polymer obtained under these condit,ions. At the optimum conditions of a contact, time of 7.1 seconds and a temperature of 585" C. (Figure 6), only 3.1.oj, polymer was formed, whereas cracking without nitrogen resulted in 9.0% polymer. Simultaneously with decreased polymer yield, a greatly increased yie'd of unsaturated esters was obtained. Thus, a t the optimum operating conditions for each method of pyrolysis (Figure 5 ) , the percent,age of unsaturates increased from 3.7 to 9.374. The unsaturated esters c.an.be cracked further under the same conditions as the diest,er to give increased yields of but.adime. The above results can be explained from the kinetic theory of gaseous reactions. Nitrogen is effective in reducing the amount of polymer formation because i t lowers the butadiene concentration in the reaction chamber and also because the presence of oxygen, which is believed to be a catalyst for the polymerization reaction, has been eliminated entirely. The increased amount of unsaturated esters was attributed to the fact that the over-all reaction is a result of t,wo consmxt>ivereactions. The cracking of t,he diacet.ate to the unsaturated a&at.e is rapid and is the normal reaction below 500" C. where no butadiene is formed. This step is t'herefore not greatly retarded by the presence of nit,rogrn. The second step of cracking t.he unvaturated ester to butadiene is much. slower and is therefore more readily ret,arded by the presence of an inert component, even at relat,ively high temperatures. The effect of the amount of nitrogen passed in was not clearly shown, since best conditions of operation were obtained with a flow of about 600 ec. per minute of nitrogen, which was kept constant for these runs. The results show t,hat varying the concentmt,ion of diacetate in the incoming stream from 50 to 17% did not, affect the results greatly since they correlate very well. CONCLUSION
a -I
w
* S
a W
5 2 7
x S
3
Pyrolysis in a n atmosphere of nitrogen greatly improved the performance of a laboratory cracking unit for converhg 2,3butylene glycol diacetat,e to 1,3-butadiene, and is recommended for commercial consideration. .4n optimum yield for one-pass unit of 83.9% conversion obtained, which probably could he raised to a t least 90% by subsequent cracking. The optimum conditions of operation, as obtained in the laboratory, are pyrolysis to 585" C. a t a contact time of 7.1 seconds, the diaceta'te being diluted by nitrogen to a concentrat,ion of about' 40 mole yoin the reaction zone. ACKNOWLEDGMENT
I
Figure 4.
Unreacted Diacetate VS. Contact T i m e (Second Quadrant) Unsaturated Acetate VI. Contact T i m e (Second Quadrant) Figure 6. Polymer Yield VI. Contact T i m e (Second Quadrant) Figure 5.
tained by difference and not by dirrut analysis. However, representative lines can be drawn through the respective points; they illuqtrate that at constant temperature the amount of polymer increases with decreasing feed rates, while a t constant contact time the polymer formation increases sharply with increasing temperature. A comparison between the amount ,of polymer formed as against the amount of unsaturated esters (particularly a t higher temperatures) shows that polymer formation is the
hppreciittion is gratefully expressed t o Schenley Research Institute and t,o A. J. Licbmann for a s s i s h w a.nd supply of rnat,erialsand equipment used in t,his work. LITERATURE CITED
(1) Hill and Isaacs, U. S. Patent 2,224,914 (Feb. 17, 1940). (2) Northern Regional Research Lab., Summary of Research on Production of Butadiene from 2J-Butylene Glycol by Pyrolysis of Diacetate, Aug., 1942. (3) Othmer, D. F., Bergen, W. S., Shlechter, N., and Bruins, P. F., IND. E N G . C H E M . , 37,890 (1945). (4) Othmer, D. F., Shlechter, N., and Koszalka, W. A., Ibid., 37, 895 (1945). (5) Pittsburgh-Des Moines Steel Co., Resszrch BUZZ.7381-B (1942). (6) Shlechter, N., Othmer, D. F., and Marshak. W. B., IND. ENG. CHEM., 37, 900 (1945).
PRESENTED aa part of the Unit Process Symposium before the Division of Industrial and Engineering Chemistry a t the 108t,h Mpetinp of the .IMERIC CHEMICAL SOCIETY in New York, N. Y.